Photo
by Istvan Kadar Photography. Image courtesy of Getty.
This dataset includes statistics about skyscrapers in the US.
Variables include the name, the city, the materials used, the height (in meters), and more.
The data was sourced from the Council on Tall Buildings and Urban Habitat (formerly known as The Skyscraper Center).
In this report, we will be investigating three predictor variables (material, floors, and height) and their relationship to the outcome variable (years to build a skyscraper).
As we can see from the boxplot and histogram, the distribution of the years it took to build a skyscraper is heavily skewed right with a five-number summary of 0, 2, 2, 3, 19 (minimum, lower quartile, median, upper quartile, and maximum respectively).
For each predictor variable, we will create a relevant graph that shows its relationship with the outcome variable and calculate any relevant bivariate statistics.
From the boxplots, it appears that the skyscrapers built from steel, concrete, and composite material have a significant amount of outliers when compared to that of skyscapers built with masonry and concrete/steel.
This may suggest that the mean of years it took for skyscapers to be built from steel, concrete and composite material could potentially be higher than that of skyscapers built with masonry, and concrete/steel materials.
| Materials | Median of Years to Build Skyscraper | Mean of Years to Build Skyscraper |
|---|---|---|
| composite | 3 | 2.40 |
| concrete | 2 | 2.25 |
| concrete/steel | 3 | 2.95 |
| masonry | 2 | 2.31 |
| steel | 2 | 2.96 |
However, when looking at the table, we can see that there is a minimal spread between both the median and mean. The median only differs by 1 and the mean only differs by a max of 0.71.
From this, we can most likely conclude that the material has no effect on the number of years it takes for a skyscraper to be built.
From the graph, we can see a majority of plots plotted within 1-4 on the x-axis and about 5-60 on the y-axis, making it hard to see a line of best fit. The correlation of the scatterplot, 0.15, is positive, but also very weak.
From this, we can most likely conclude that the number of floors on skyscraper has no effect on the number of years it takes for a skyscraper to be built.
Compared to the last scatterplot, this one seems to show a stronger correlation than that between years and number of floors. However, similarly to the last one, the outliers make it difficult to make out a line of best fit. Although the correlation of 0.2 is stronger than that between years and number of floors, it is still quite too weak to establish a strong correlation between the two variables.
Although it is reasonable to believe that the three separate variables (material, number of floors, and height) would have some sort of effect on the years it would take for a skyscraper to be built, we have found that there is not strong enough evidence to conclude that.
Material used: Difference in both mean and median are too small to be considered significant.
Number of floors: Correlation is too weak to be considered significant.
Height (in meters): Correlation is too weak to be considered significant.
Some other areas to explore that could effect the time it takes to build a skyscraper could include: construction company, total cost of project, climate, total number of workers, etc.
Dataset sourced from here.